中文摘要：

提出了一种基于Kalman滤波的数字水印算法,利用图像融合的思想和Kalman滤波对数字水印进行了研究。嵌入水印的过程可以视为宿主图像和水印图像的图像融合的过程。假设嵌水印图像作为目标状态,在图像融合过程中,采用Kal-man滤波对观测状态进行最优估计。根据宿主图像和水印图像进行建模,得到观测方程和状态方程,在估计误差方差最小意义下给出最优估计结果。最后基于交叉熵和互信息指标对所提出算法的性能进行了评价。实验结果表明,基于Kalman滤波的数字水印算法具有较好的鲁棒性和不可感知性。

英文摘要：

A digital watermarking algorithm based on Kalman filter is proposed.The watermarking algorithm is researched according to the image fusion and Kalman filtering.The process of watermarking can be viewed as a process of image fusion from the original image and the watermark image.Assuming that the watermarked image is the steady-state,Kalman filter is used as an optimal estimation algorithm in the process of image fusion.The math model is built according to the watermark image and the original image.Then the state equation and the corresponding measurement equation are built.The optimal estimation is received in the case of minimum estimation error variance.The Cross Entropy and the Mutual Information are used to evaluate the performance of proposed watermark algorithm.Experimental results show that the proposed algorithm has a good performance both in robustness and invisibility.